RESIDUAL STRESS AND TEXTURE DUE TO COLD AND HOT EXTRUSION PROCESSES

The residual stress state and the texture of cold forward extruded full and hollow steel bodies as well as a hot extruded A1Si25Cu4Mgl tube are studied by X-ray, high energy synchrotron and neutron diffraction. The experimental results reveal that all samples are fibre textured and that there are characteristic distributions of the residual stresses vs. sample diameter. In case of the cold forward extruded samples at low degrees of natural strain, the rod kernel is under compressive residual stresses which are balanced by tensile residual stresses in the outer part of the sample. In contrast to this, the outer part of the hot extruded sample is under compressive macroscopic stresses which are balanced by tensile macroscopic residual stresses in the inner part of the sample.


INTRODUCTION
The strong plastic deformation and the macroscopic deformation gradient during the extrusion process lead to the formation of grain deformation, texture and residual stresses. The microstructure, the texture and especially the residual stress state established during an extrusion process are of great practical interest regarding the static and dynamic mechanical properties of the material as well as the workpiece's resistance to shock and fatigue failure. Therefore, the texture and the residual stress state of cold forward extruded steel samples and 292 A. PYZALLA AND W. REIMERS the hot forward extruded A1Si-alloy A1Si25CU4Mg1 are investigated by X-ray, neutron and high energy synchrotron diffraction. The use of non-destructive methods is most profitable here, since considerably differing results have been reported after the determination of the residual stresses in extruded samples by destructive methods (Frisch and Thomsen, 1957;Moore and Evans, 1958), while residual stress analysis by neutron diffraction (Modlen et al., 1992;Genzel et al., 1996) revealed systematic relations between process parameters and the resulting residual stress state of the specimens. Furthermore, in case of the hot forward extruded multiphase A1Si25Cu4Mgl-alloy, phase specific residual microstresses, which only can be determined by diffraction methods, arise due to the deformation process as well as due to cooling. of Stuttgart, Germany. The degree of natural strain qo was varied between qo 0.5, 1.2 and 1.6 in case of full forward extruded samples and between qo 0.5 and 1.2 in case of the hollow samples. The A1Si25-Cu4Mgl tubes were hot forward extruded at the "Forschungszentrum Strangpressen" of the Technical University in Berlin, Germany. The extrusion process was performed indirectly using a moving mandrel. The temperature of the workpiece was 360C app., the degree of natural strain is o 2.8.

Texture Characterisation and Residual Stress Analysis
For texture analysis an X-ray diffractometer with an Eulerian cradle was employed. For neutron diffraction the instrument E3 of the BERII reactor at the Hahn-Meitner-Institut, Berlin, Germany was used. A wavelength A 0.13880 nm (Cu-monochromator, 220) was chosen. In case of the steel samples this covers the reflections 110/220, 211 and 200 within the experimentally acw.essible range 20.. 40-120 of the multichannd counter used for radiation detection. The diffracting volume was limited by conical cadmium slits to 2 x 2 x 2 mm3. In order to account for the texture and the plastic anisotropy, the difference between the axial and the radial residual stress component was evaluated from the slope of the d vs. sin 2 curves. Whereas the d vs. sin 2 b curves were linear in case of the hollow samples, non-linear d vs. sin 2 b curves were obtained in the rod kernel of the full forward extruded sample and thus were evaluated using appropriate methods (Hauk and Sesemann, 1976;1985;Hauk et al., 1990). For the interplanar lattice spacing do of the Si particles in the A1Si25Cu4Mgl, the d-value of pure Si according to the JCPDS was used. The do of the Al-alloy matrix was calculated from equilibrium conditions.
Residual stress analysis using high energy synchrotron radiation was performed at the beam line ID 15A of the ESRF Grenoble. Due to the high photon flux and the parallel beam the volume element in this case could be restricted to a small parallelepiped of 1.65mm x 0.145mm.
The do value necessary for the determination of the three-dimensional residual stress state was calculated as an average of the d-values obtained for the different reflections and volume elements ('randomwalk-method').

Cold Forward Extrusion
The microstructure of a cold forward extruded steel sample reveals severe grain elongation caused by the plastic deformation during RESIDUAL STRESS AND TEXTURE 295 the extrusion process (Fig. 3). The plastic deformation also leads to the development of a (110)-fibre texture (Fig. 4), characteristic for bcc steels (Wassermann and Grewen, 1962). This fibre texture is more pronounced in the rod kernel than in the outer part of the samples.  Figure 5 shows the distribution of the axial residual stress component vs. sample diameter of a full forward extruded specimen. Neutron diffraction and synchrotron diffraction reveal in very good agreement, that in the inner part of the specimen the residual stresses in radial direction rrr, hoop **, and axial direction zz, are compressive. These compressive residual stresses are balanced by tensile residual stresses in the outer part of the sample (Modlen et al., 1992). At the surface of the sample again compressive stresses in hoop and axial direction were obtained by X-ray diffraction. This residual stress distribution, which is characteristic for hollow as well as full forward extruded samples, can be linked to the deformation process during the cold forward extrusion (Tekkaya, 1986). Due to deformation obstruction at the shoulder of the die, the material flow at the outer surface is slower and more inhomogeneous than in the inner part of the specimen. Therefore, within the inner part of the samples the grains are homogeneously stretched whereas at the outer surface of the samples the grains are first compressed and stretched later, while passing the transient radius of the die. Thus, tensile residual stresses remain in the outer part, while the inner part of the samples is under compressive residual stresses. principle of residual stress formation holds for the hollow extruded samples. Therefore, the residual stress distributions obtained versus the diameter of a hollow sample (Fig. 6) is quite similar to the one of a full forward extruded sample.
Due to the weak texture and weak plastic anisotropy, which can be attributed to the low degree of natural strain of 0.5, the axial stress component obtained for the two reflections 200 and 211 are in good agreement. Furthermore, the experimental results are in good qualitative agreement with Finite-Element-Calculations (Tekkaya, 1986) of the axial residual stress component (Fig. 7) in a sample that is slightly thicker but extruded with the same degree of natural strain.

Hot Extrusion
Whereas the steel C15 here can be regarded as a single-phase material, the alloy AISi25Cu4Mgl is a multi-phase material containing large Si particles and smaller particles of an intermetallic phase (Fig. 8). Due to the low amount of intermetallic phase, it is not visible in the diffraction diagrams.
The texture of the hot extruded AISi25Cu4Mgl alloy is a two-folded fibre texture with orientations (100) and (111) as fibre axis (Fig. 9) which is typical for extruded fcc materials (Wassermann and Grewen, 1962). The comparatively weak texture with regard to the degree of natural strain of the sample, according to Wassermann and Grewen (1962), can be attributed to the presence of the hard Si particles. In opposition to the residual stress distribution in the cold extruded full and hollow samples the macroscopic residual stresses in axial direction are tensile within the inner part of the hot extruded sample, while the outer part of the hot extruded sample contains the balancing compressive residual stresses (Fig. 10).
The difference in the residual stress distribution vs. sample diameter on the one hand can be attributed to the larger degree of natural strain of o 2.8 possible in hot extrusion, since tensile residual stresses also were observed in case of a cold extruded sample with a comparatively high degree of natural strain of qo 1.6 (Pyzalla et al., 1996;Pyzalla and Reimers, 1997). On the other hand, due to the macroscopic temperature gradient during cooling from the extrusion temperature, also compressive residual stresses arise near the outer surface of the tubes while balancing tensile residual stresses are induced at smaller sample diameters. Since the Al-alloy matrix of AISi25Cu4Mgl has a substantially larger thermal expansion coefficient than the Si particles, the cooling also produces phase specific residual microstresses. These are

CONCLUSIONS
The microstructure, the texture and the residual stress state of cold forward extruded and hot forward extruded samples were investigated by X-ray, high energy synchrotron and neutron diffraction. The experiments reveal characteristic residual stress distributions vs. sample diameter. In case of the cold forward extruded samples compressive residual macrostresses are present in the rod kernel which are balanced by tensile residual stresses at larger sample diameters. In opposition, the hot extruded sample contains tensile residual stresses in the inner and compressive residual stresses in the outer part. From these experimental results it can be concluded that the residual stress distribution over the sample diameter can be largely influenced by the degree of natural strain and the extrusion temperature and, that thus a favourable residual stress state may be established using optimised process parameters.